At Perplexity, I developed autonomous AI workflows with Python, LangGraph, and GPT-based reasoning agents for complex, multi-step tasks used by over 120,000 monthly users. I also improved task completion accuracy by 31% through agent memory, retrieval augmentation, and structured reasoning.
I built multi-agent architectures and RAG pipelines, and reduced response latency by 42% through execution graph and processing optimizations. I also lowered LLM inference and operational costs by 28% with model routing, caching, and token optimization.
At Accenture, I developed machine learning pipelines and MLOps workflows supporting over 12,000 internal business users. I improved prediction accuracy by 18% and earned certifications in AWS machine learning, AI, Databricks machine learning, and Kubernetes application development.
